| With the advent of the digital age,broadband wireless access technology has developed rapidly,and mobile technology has been widely used.LBS has an increasingly urgent need for precise positioning.Smart terminals support Wi Fi and Bluetooth functions,prompting indoor positioning technology to further tilt toward Wi Fi and Bluetooth technology.The research on Wi Fi positioning,Bluetooth positioning and the positioning technology of the integration of the two has great application value.Aiming at the problems of Wi Fi and Bluetooth receiving signal fluctuations,large positioning errors and insufficient stability of the single positioning technology,this paper designs a strong signal WKNN positioning algorithm based on Wi Fi fingerprint range matching and an area weighted positioning based on Bluetooth fingerprints.Algorithm,and on this basis,design a fusion positioning algorithm that uses Bluetooth positioning to assist Wi Fi fingerprint positioning.The research content of this paper is as follows:(1)A WKNN location algorithm for strong signal clusters based on Wi Fi fingerprint range matching is designed.First,during the offline establishment of a fingerprint database,the maximum value of the received signal strength(RSS)and the minimum value of the RSS are additionally stored for each fingerprint record.Then,in the online matching stage,the range matching algorithm is executed according to the RSS of the input Wi Fi signal vector to obtain the confidence of each fingerprint,and fingerprints with a confidence of 0are removed.Finally,the input Wi Fi signal vector is filtered according to the RSS threshold of the strong signal Wi Fi to obtain the Wi Fi signal sub-vector,and the WKNN algorithm with fingerprint confidence as the weight distribution standard is executed.Experimental results show that compared with the distance-weighted WKNN algorithm,the probability of positioning error within 3m increases from 77.5% to 83.3%.(2)Design an area weighted positioning algorithm based on Bluetooth fingerprint.First,in the offline phase,the positioning scene is divided into regions according to the deployment position of Bluetooth,the region number is set,and the fingerprints of these regions are additionally collected to establish a Bluetooth region fingerprint library.Then,in the online phase,the KNN algorithm is executed by inputting the i Beacon signal vector and all the Bluetooth reference point fingerprint vectors to obtain the candidate Bluetooth fingerprint.Then,obtain the regional fingerprint corresponding to the region number of the candidate fingerprint,and calculate the cosine similarity between the input i Beacon signal vector and the regional fingerprint vector,and finally use the cosine similarity value as the standard to assign fingerprint weights and execute the WKNN algorithm.Experimental results show that compared with the distance-weighted WKNN algorithm,the probability of positioning error within 4m increases from 81.7% to 85%.(3)A fusion positioning algorithm of Wi Fi and Bluetooth is designed.First,based on the Bluetooth fingerprint area weighted positioning algorithm,an algorithm is designed to determine the possible area number set of the points to be measured.The algorithm determines the circle range with the Bluetooth fingerprint positioning result as the center and the maximum positioning error as the radius,and traverses The fingerprint in the range circle obtains the possible area number set of the point to be measured.Then,filter the original Wi Fi fingerprint database according to these area number sets to obtain a new Wi Fi fingerprint database.Finally,according to the Wi Fi signal vector of the point to be measured and the new Wi Fi fingerprint library,the WKNN positioning algorithm with strong signal matching the Wi Fi fingerprint range is executed to obtain the fusion positioning result.Experimental results show that,compared with the single Wi Fi positioning technology,the probability of positioning error within 2m increases from53.3% to 69.2%. |
PDF Full Text Request